This double-blind, placebo-controlled, study will attempt to evaluate the relative efficacy of two tricyclic antidepressants (doxepin and imipramine), a major tranquilizer (thioridazine) and a minor tranquilizer (chlordiazepoxide) in the treatment of non-psychotic outpatients characterized as primarily anxious or as primarily depressed. After careful screening and a two-week placebo washout period, patients will be randomly assigned to the five treatment conditions for a period of eight weeks. Three collaborating clinics will each attempt to complete 200 patients; half primarily anxious and half primarily depressed. Symptomatic volunteers will be used in one setting and regular clinical patients in the other two settings. An attempt will be made to identify reliable predictors of differential drug response in order to identify the right drug for the right patient. Methodological issues regarding dropout, medication compliance, and the relative sensitivity of doctor and patient ratings of therapeutic change will be explored. This proposal represents a major attempt to provide concurrently obtained data on three classes of psychotropic drugs thought to be effective in the treatment of a broad spectrum of anxious and depressed outpatients.